|Chenshen Wu||Computer Vision Center|
|Luis Herranz||Computer Vision Center|
|Xialei Liu||Computer Vision Center|
|Yaxing Wang||Centre de Visió per Computador (CVC)|
|Joost Van De Weijer||Computer Vision Center Barcelona|
|Bogdan Raducanu||Computer Vision Center|
Previous works on sequential learning address the problem of forgetting in discriminative models.Addressing this problem, the authors propose Memory Replay GANs (MeRGANs), a conditional GAN framework that integrates a memory replay generator.
Previous works on sequential learning address the problem of forgetting in discriminative models. In this paper we consider the case of generative models. In particular, we investigate generative adversarial networks (GANs) in the task of learning new categories in a sequential fashion. We first show that sequential fine tuning renders the network unable to properly generate images from previous categories (i.e. forgetting). Addressing this problem, we propose Memory Replay GANs (MeRGANs), a conditional GAN framework that integrates a memory replay generator. We study two methods to prevent forgetting by leveraging these replays, namely joint training with replay and replay alignment. Qualitative and quantitative experimental results in MNIST, SVHN and LSUN datasets show that our memory replay approach can generate competitive images while significantly mitigating the forgetting of previous categories.